Care Agent Call Classification

1. A computer-based method of analyzing call center data comprising: collecting, by one or more computers, reports from a plurality of customer care agents, wherein each report comprises logged data of incoming customer calls; compiling, by the one or more computers, the reports from the plurality of customer care agents into an aggregate data set; determining, by the one or more computers, a number of total incoming calls in the aggregate data set; dividing, by the one or more computers, the aggregate data set into a plurality of categories based on the incoming calls; determining, by the one or more computers, an aggregate distribution for each of the plurality of categories, each aggregate distribution corresponding to a ratio of a number of incoming calls reported by the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the aggregate data set; compiling, by the one or more computers, reports from one of the plurality of customer care agents into an individual data set; determining, by the one or more computers, a number of total incoming calls in the individual data set; dividing, by the one or more computers, the individual data set into the plurality of categories; determining, by the one or more computers, an individual distribution for each of the plurality of categories, each individual distribution corresponding to a ratio of a number of incoming calls reported by the one of the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the individual data set for one or more of the plurality of categories, comparing, by the one or more computers, the individual distribution to the aggregate distribution; for the one or more of the plurality of categories, determining, by the one or more computers, a deviation of the individual distribution from the aggregate distribution, based on the comparison; for the one or more plurality of categories, determining, by the one or more computers, whether the deviation exceeds a threshold value; and taking action when determining that the deviation exceeds the threshold value.

2. The method of claim 1 , wherein the deviation is determined as one of an absolute deviation and a standard deviation.

3. The method of claim 1 , wherein the action comprises one or more of managing call center personnel, modifying call center personnel data logging procedures, and managing customer service.

4. The method of claim 3 , wherein managing call center personnel comprises one or more of training call center personnel, hiring call center personnel, and terminating call center personnel, and monitoring call center personnel.

5. The method of claim 1 , further comprising removing suspect data from the collected reports, wherein suspect data comprise one or more of duplicative data, low-volume data, high-frequency data, and fraudulent data.

6. The method of claim 5 , wherein low-volume data comprise a number of calls to a given customer care agent that is below a pre-determined number of calls, in a given period of time.

7. The method of claim 5 , wherein high-frequency data comprise calls with timestamps that are too short to represent an actual incoming customer call.

8. The method of claim 1 , wherein comparing the individual distribution to the aggregate distribution provides a confidence level.

9. A computer-based method of analyzing call center data comprising: collecting, by one or more computers, data of incoming customer calls from a plurality of customer care agents to form an aggregate data set; determining, by the one or more computers, a number of total incoming calls in the aggregate data set; dividing, by the one or more computers, the aggregate data set into a plurality of categories based on types of incoming customer calls in the aggregate data set; determining, by the one or more computers, an aggregate distribution for each of the plurality of categories, each aggregate distribution corresponding to a ratio of a number of incoming calls reported by the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the aggregate data set; collecting, by the one or more computers, data of incoming customer calls from one of the plurality of customer care agents to form an individual data set; determining, by the one or more computers, a number of total incoming calls in the individual data set; dividing, by the one or more computers, the individual data set into the plurality of categories based on types of incoming customer calls in the individual data set; determining, by the one or more computers, an individual distribution for each of the plurality of categories, each individual distribution corresponding to a ratio of a number of incoming calls reported by the one of the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the individual data set; for one or more of the plurality of categories, comparing, by the one or more computers, the individual distribution to the aggregate distribution; for the one or more of the plurality of categories, determining, by the one or more computers, a deviation of the individual distribution from the aggregate distribution, based on the comparison; for the one or more of the plurality of categories, determining, by the one or more computers, whether the deviation exceeds a threshold value; removing at least a portion of the individual data set from the aggregate data set to form a, groomed data set when the deviation exceeds the threshold value; and identifying valid customer service issues based on analyzing the groomed data set distribution within the first category.

10. The method of claim 9 , further comprising removing suspect data from the collected data, wherein suspect data comprises one or more of duplicative data, low-volume data, high-frequency data, and fraudulent data.

11. The method of claim 9 , wherein the aggregate distribution corresponds to a ratio of an average number of incoming calls reported by the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the aggregate data set, and wherein the individual distribution corresponds to a ratio of an average number of incoming calls reported by the one of the plurality of customer call agents as being associated with the corresponding category to the total incoming calls in the individual data set.

12. A computer-based method of analyzing call center data comprising: collecting, by one or more computers, data of incoming customer calls from a plurality of customer call agents to form an aggregate data set; determining, by the one or more computers, a number of total incoming calls in the aggregate data set; dividing, by the one or more computers, the aggregate data set into a plurality of categories based on types of incoming customer calls in the aggregate data set; determining, by the one or more computers, an aggregate distribution for each of the plurality of categories, each aggregate distribution corresponding to a ratio of a number of incoming calls reported by the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the aggregate data set; collecting, by the one or more computers, data of incoming customer calls from one of the plurality of customer care agents to form an individual data set; determining, by the one or more computers, a number of total incoming calls in the individual data set; dividing, by the one or more computers, the individual data set into the plurality of categories based on types of incoming customer calls in the individual data set; determining, by the one or more computers, an individual distribution for each of the plurality of categories, each individual distribution corresponding to a ratio of a number of incoming calls reported by the one of the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the individual data set; for one or more of the plurality of categories, comparing, by the one or more computers, the individual distribution to the aggregate distribution; for the one or more of the plurality of categories, determining, by the one or more computers, a deviation of the individual distribution from the aggregate distribution, based on the comparison; setting, by the one or more computers, alarm levels at one or more predetermined threshold levels for deviation values to indicate areas of and for the one or more of the plurality of categories, determining, by the one or more computers, whether the determined deviation exceeds at least one of the one or more predetermined threshold levels.

13. The method of claim 12 , wherein the alarm levels are set based on a pre-determined number of customer call types within the plurality of categories.

14. The method of claim 13 , wherein customer call types are one or more of dropped calls, billing problems, and hardware problems.

15. The method of claim 12 , further comprising removing suspect data from the collected data, wherein suspect data comprise one or more of duplicative data, low-volume data, high-frequency data, and fraudulent data.

16. The method of claim 15 , wherein duplicative data comprise incoming call data to a given customer care agent from similar phone numbers.

17. The method of claim 15 , wherein low-volume data comprise the number of calls to a given customer care agent being below a pre-determined number of calls in a given period.

18. The method of claim 15 , wherein high-frequency data comprise calls with durations determined to be too short to represent an actual incoming customer call.

19. The method of claim 1 , wherein taking an action comprises removing at least one report collected from the one of the plurality of customer care agents from the aggregate data set.

20. The method of claim 9 , further comprising: dividing, by the one or more computers, the groomed data set into the plurality of categories based on types of incoming customer calls in the groomed data set; and determining, by the one or more computers, a groomed distribution for each of the plurality of categories, each groomed distribution corresponding to a ratio of a number of incoming calls in the groomed data set reported by the plurality of customer call agents as being associated with the corresponding category to the number of total incoming calls in the groomed data set.